Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
SUBMILLIMETER OBSERVATIONS OF GIANT MOLECULAR CLOUDS IN THE LARGE MAGELLANIC CLOUD: TEMPERATURE AND DENSITY AS DETERMINED FROM J = 3Y2 AND J = 1Y0 TRANSITIONS OF CO Tetsuhiro Minamidani,1, 2 Norikazu Mizuno,1 Yoji Mizuno,1 Akiko Kawamura,1 Toshikazu Onishi,1 Tetsuo Hasegawa,3 Ken’ichi Tatematsu,3 Masafumi Ikeda,4 Yoshiaki Moriguchi,1 Nobuyuki Yamaguchi,3 Jürgen Ott,5, 6 Tony Wong,7, 8 Erik Muller,7 Jorge L. Pineda,9 Annie Hughes,7, 10 Lister Staveley-Smith,11 Ulrich Klein,9 Akira Mizuno,1, 12 Silvana Nikolić,13, 14 Roy S. Booth,13, 15 Arto Heikkilä,13 Lars-8ke Nyman,16 Mikael Lerner,16 Guido Garay,17 Sungeun Kim,18 Motosuji Fujishita,1 Tokuichi Kawase,1 Monicá Rubio,17 and Yasuo Fukui1 ABSTRACT We have carried out submillimeter 12 CO(J ¼ 3Y2) observations of six giant molecular clouds (GMCs) in the Large Magellanic Cloud (LMC) with the ASTE 10 m submillimeter telescope at a spatial resolution of 5 pc and very high sensitivity. We have identified 32 molecular clumps in the GMCs and revealed significant details of the warm and dense molecular gas with n(H2 ) 10 3 Y10 5 cm3 and Tkin 60 K. These data are combined with 12 CO(J ¼ 1Y0) and 13CO(J ¼ 1Y0) results and compared with LVG calculations. The results indicate that clumps that we detected are distributed continuously from cool (10Y30 K) to warm (k30Y200 K), and warm clumps are distributed from less dense (10 3 cm3) to dense (10 3.5 Y10 5 cm3). We found that the ratio of 12 CO(J ¼ 3Y2) to 12 CO(J ¼ 1Y0) emission is sensitive to and is well correlated with the local H flux. We infer that differences of clump properties represent an evolutionary sequence of GMCs in terms of density increase leading to star formation. Type I and II GMCs (starless GMCs and GMCs with H ii regions only, respectively) are at the young phase of star formation where density does not yet become high enough to show active star formation, and Type III GMCs (GMCs with H ii regions and young star clusters) represent the later phase where the average density is increased and the GMCs are forming massive stars. The high kinetic temperature correlated with H flux suggests that FUV heating is dominant in the molecular gas of the LMC. Subject headingg s: galaxies: individual ( LMC) — ISM: clouds — ISM: molecules — Magellanic Clouds — radio lines: ISM — submillimeter Online material: color figures 1. INTRODUCTION It is of fundamental importance in astronomy to understand the evolution of galaxies. Since a major constituent of galaxies is stars, the formation of stars is a fundamental process in galactic evolution. The properties of stars characterize the basic contents of galaxies and their time evolution. We understand from studies of the Milky Way that giant molecular clouds (GMCs), whose mass ranges from 10 5 to 107 M, are the principal sites of star formation and that this perhaps holds true in other galaxies as well. We also recognize that the GMC properties (e.g., LCO Yline width relation, index of mass spectrum) are similar among five galaxies 1 Department of Astrophysics, Nagoya University, Furo-cho, Chikusa-ku, Nagoya 464-8602, Japan. 2 Current address: Department of Physics, Faculty of Science, Hokkaido University, N10W8, Kita-ku, Sapporo 060-0810, Japan. 3 National Astronomical Observatory of Japan, Mitaka, Tokyo 181-8588, Japan. 4 Research Center for the Early Universe and Department of Physics, University of Tokyo, Tokyo 113-0033, Japan. 5 National Radio Astronomy Observatory, 520 Edgemont Road, Charlottesville, VA 22903-2475. 6 Jansky Fellow of the National Radio Astronomy Observatory. 7 CSIRO Australia Telescope National Facility, P.O. Box 76, Epping, NSW 1710, Australia. 8 Current address: Department of Astronomy, MC 221, University of Illinois, Urbana, IL 61801. 9 Radioastronomisches Institut der Universität Bonn, Auf dem Hügel 71, 53121 Bonn, Germany. in the Local Group according to the spatially resolved studies ( Blitz et al. 2007). This supports the idea that studies of GMCs will be useful in understanding the fundamentals of galactic evolution through the formation and evolution of GMCs and star formation therein. Observational studies of GMCs have been most effectively made by the millimeter interstellar carbon monoxide emission line at 2.6 mm, which allows us to probe molecular gas whose density is greater than 100 cm3. We note that the most abundant species, molecular hydrogen, does not have appropriate line emissions in the millimeter and submillimeter region due to its zero permanent electric dipole moment and large separation between 10 Center for Supercomputing and Astrophysics, Swinburne University of Technology, Hawthorn, VIC 3122, Australia. 11 School of Physics, M013, University of Western Australia, 35 Stirling Highway, Crawley, WA 6009, Australia. 12 Current address: Solar-Terrestrial Environment Laboratory, Nagoya University, Furo-cho, Chikusa-ku, Nagoya 464-8601, Japan. 13 Onsala Space Observatory, 439-92 Onsala, Sweden. 14 Current address: Departament de Astronomı́a, Universidad de Chile, Casilla 36-D, Santiago, Chile. 15 Current address: Hartebeesthoek Radio Astronomy Observatory, P.O. Box 443, Krugersdorp 1740, South Africa. 16 European Southern Observatory, Casilla 19001, Santiago 19, Chile. 17 Departament de Astronomı́a, Universidad de Chile, Casilla 36-D, Santiago, Chile. 18 Astronomy and Space Science Department, Sejong University, 98 Kwangjin-gu, Kunja-dong, Seoul 143-747, Korea. the lowest energy levels, which are not excited significantly in the typical physical conditions of molecular clouds. Recent advances in submillimeter observations have allowed us to determine physical parameters of molecular clouds over much larger ranges than in the millimeter region by comparing line intensities between different transitions. These submillimeter studies were initiated by the Swedish-ESO Submillimetre Telescope (SEST) 15 m telescope in Chile followed by instruments in Mauna Kea, Hawaii, and in the Swiss Alps at an altitude range from 3700 to 4200 m, including the Caltech Submillimeter Observatory (CSO) 10 m, James Clerk Maxwell Telescope (JCMT) 15 m, and KOSMA 3 m telescopes, and the Antarctic Submillimeter Telescope and Remote Observatory (AST/ RO) 1.6 m telescope in Antarctica. Subsequently, in the 2000s, the development of new instruments at an altitude of 5000 m in Atacama in northern Chile resulted in a superior capability because of the high altitude and dry characteristics of the site. The instruments installed in Atacama include the Atacama Submillimeter Telescope Experiment (ASTE) 10 m, APEX 12 m, and NANTEN2 4 m telescopes. All these instruments are beginning to take new molecular data with significantly better quality than before in terms of noise level, as well as angular resolution. It is also noteworthy that the current frequency coverage extends as high as the 800 GHz band and even the terahertz region. Among nearby galaxies that we can observe at reasonably high spatial resolutions, the Large and Small Magellanic Clouds offer us a unique opportunity to achieve the highest resolutions due to their unrivaled closeness, 50Y60 kpc. In particular, the Large Magellanic Cloud (LMC) is actively forming stars in clusters and is an ideal laboratory for us to study star formation, particularly massive star formation in star clusters. In the LMC, the metallicity is a factor of 3 lower than in the solar neighborhood (Dufour et al. 1982; Dufour 1984; Rolleston et al. 2002). In addition, the visual extinctions are lower and the FUV field is stronger in the LMC than in the Milky Way (Israel et al. 1986), characterizing the initial conditions of star formation. The first spatially resolved complete sample of GMCs in a single galaxy has been obtained toward the whole LMC with the NANTEN 4 m telescope in 2.6 mm CO emission at 40 pc resolution (Fukui et al. 1999, 2001, 2007; Mizuno et al. 2001). These studies revealed the three types of GMCs in terms of star formation activities; type I is starless, type II is with H ii regions only, and type III is associated with active star formation indicated by huge H ii regions and young star clusters, where the stars identified are only O stars due to the sensitivity limitation. It also revealed that the lifetime of a GMC is as short as 30 Myr (Fukui 2007; Kawamura et al. 2007; A. Kawamura et al. 2008, in preparation). These previous studies naturally place the LMC as one of the prime targets for submillimeter studies to derive the physical parameters of GMCs. Another aspect that deserves our attention is that very young, rich stellar clusters are forming in the LMC. These are so-called populous clusters, which are very rare in the Milky Way and resemble globular clusters formed in the primeval Milky Way. The open clusters forming in the Milky Way are small in the number of stars and loose in spatial distribution. Along with the low metallicity of the LMC, it is an interesting possibility to use molecular data to investigate the formation mechanism of super star clusters at the molecular cloud stage. In the past, there have been some studies that used the higher transitions (J ¼ 2Y1, J ¼ 3Y2, J ¼ 4Y3, J ¼ 7Y6) of CO spectra of the molecular clouds in the LMC (e.g., Sorai et al. 2001; Johansson et al. 1998; Heikkilä et al. 1999; Bolatto et al. 2005; Israel et al. 2003; Kim et al. 2004; Kim 2006). These studies suggest that the molecular gas may be warmer and/or denser than in the Milky Way. Johansson et al. (1998) used the SEST 15 m telescope to observe the central part of the 30 Doradus Nebula (rms 0.2 K at 0.5 km s1 velocity resolution for J ¼ 1Y0 and rms 1.0 K at 0.5 km s1 velocity resolution for J ¼ 3Y2) and the southern H ii regions N158C, N159, and N160 with a few prominent CO clouds in the J ¼ 2Y1 and J ¼ 3Y2 transitions of CO. They find that the kinetic temperatures are 10Y80 K and the highest temperature is toward 30 Dor. The smallest beam size and grid spacing are 1500 and 1100 , respectively, in the J ¼ 3Y2 emission. Heikkilä et al. (1999) used SEST to observe the J ¼ 3Y2 transition of CO in N159 and 30 Doradus, as well as other rarer molecular species. This study aimed at obtaining chemical abundances, while it also provides more information on cloud temperature, etc., from CO(J ¼ 3Y2) data. The kinetic temperatures that they derived are 50 K in 30 Dor-10, 15 K in 30 Dor-27, and 20Y25 K in N159W and N160. Bolatto et al. (2005) employed the AST/ RO to observe the 12 CO(J ¼ 4Y3) transition at 461 GHz with a 10900 beam. They observed nine regions in the LMC at 6 0 ; 6 0 field, all with H ii regions, and derived kinetic temperatures from a comparison between the CO(J ¼ 4Y3) and (J ¼ 1Y0) transitions. N48, N55A, N79, N83A, N113, N159W, N167, N214C, and LIRL 648 are included. They derive temperatures of 100 Y 300 K and note a trend that higher temperatures occur in moderatedensity regions, 100Y1000 cm3, and the lower temperatures in much denser regions, 10 4 Y10 5 cm3. These studies were preceded by a suggestion that significant amounts of warm molecular gas may exist in the LMC (Israel et al. 2003). Kim et al. (2004) also made similar observations toward an H ii region, N44, and suggest very dense gas of 105 cm3. Most recently, Kim (2006) derived Tkin ¼ 100 K and n 10 4:3 cm3 for 30 Dor from the intensity ratios of 12 CO(J ¼ 7Y6) to 12 CO(J ¼ 4Y3) and 12 CO(J ¼ 1Y0) to 13CO(J ¼ 1Y0). In the present study, we aim to obtain submillimeter molecular data at better signal-to-noise ratios than in the previous studies to make estimates of temperatures and densities over a large sample in the LMC. We combine the 12 CO(J ¼ 3Y2) data obtained with the ASTE telescope and CO(J ¼ 1Y0) data obtained with the SEST and Mopra telescopes. In order to make reasonable comparisons between the two transitions, J ¼ 3Y2 and J ¼ 1Y0, we smooth the ASTE results (2200 beam) to the same resolution as the SEST data (4500 ) and use large velocity gradient ( LVG) calculations to estimate density and temperature. We also employ the 13 CO(J ¼ 1Y0) data, where available, to place constraints on the physical parameters. This paper is organized as follows: x 2 describes the observations. Sections 3 and 4 show the results and data analysis, respectively. In x 5, we discuss the physical properties of clumps and evolutional sequence of GMCs. Finally, we present a summary in x 6. 2. OBSERVATIONS 2.1. Selection of GMCs The present targets were chosen from the NANTEN catalog of CO(J ¼ 1Y0) GMCs compiled by Fukui et al. (2007). This catalog is based on the second survey, with a factor of 2 higher sensitivity than the first survey (Mizuno et al. 2001). The catalog includes 272 CO clouds, 230 of which are detected at three or more observed positions, and they are classified into the three types: 56 (24.3%) type I (starless) GMCs, 120 (52.2%) type II GMCs (those with H ii regions only), and 54 (23.5%) type III GMCs (those with H ii regions and young star clusters), where ‘‘stars’’ 12 In the present study, we observe GMCs in the southeast region of the LMC, which contains 30 Doradus, the largest and most massive H ii region in the Local Group. We observe the molecular ridge extending southward from 30 Doradus and the ‘‘CO Arc’’ along the southeastern optical edge ( Fukui et al. 1999). Three type III GMCs, LMC N J05386904 (the 30 Dor region), LMC N J05407008 (the N159 region and the N171 region), and LMC N J05307106 (the N206 region), are selected as the principal targets, and two type II GMCs, LMC N J05446923 (the N166 region) and LMC N J05327114 (the N206D region), and a type I GMC, LMC N J05477041 (the GMC 225 region), are included for reference. The locations of the observed GMCs and regions are shown in Figure 1, and their coordinates and the data used in this paper are summarized in Table 1. Hereafter, the region names, which are in the parentheses above or column (4) of Table 1, are used to identify the regions. 2.2. 12 CO(J ¼ 3Y2) Fig. 1.— CO velocity-integrated intensity map ( Fukui et al. 2001, 2007) overlaid on ESO(R) image. The observed seven regions are indicated by white circles. [See the electronic edition of the Supplement for a color version of this figure.] refer only to O stars due to the limited sensitivity of existing observations (A. Kawamura et al. 2008, in preparation; Fukui 2007; Kawamura et al. 2007; Blitz et al. 2007). Among these GMCs, we mainly focus on type III GMCs whose 12 CO(J ¼ 3Y2) intensities are expected to be strong due to the highly excited conditions. In addition, relatively high resolution 12 CO and 13 CO J ¼ 1Y0 data, observed with the SEST 15 m or Mopra 22 m telescopes, are collected to derive physical properties of molecular clouds through the LVG analysis at a 10 pc scale. Observations of the 12 CO(J ¼ 3Y2) transition at 345 GHz were made with the ASTE telescope at Pampa la Bola in Chile ( Ezawa et al. 2004) in 2004 October. The half-power beam width was measured to be 2200 at 345 GHz by observing the planets. This corresponds to 5.3 pc at the distance of the LMC, 50 kpc. The telescope was equipped with a single ‘‘cartridge-type’’ SIS receiver, sensitive from 324 to 384 GHz, which is of a similar design to that for ALMA (Kohno 2005). The spectrometer was an XF-type digital autocorrelator (Sorai et al. 2000) and was used in the wideband mode, which has a bandwidth of 512 MHz with 1024 channels. The spectrometer provided a velocity coverage and resolution of 450 and 0.44 km s1, respectively, at 345 GHz. We observed six GMCs (seven regions) in the Large Magellanic Cloud as shown in Figure 1 and listed in Table 1. These observations were carried out by position switching at a grid spacing of 2000 or 3000 for the entire clouds and of 1000 or 1500 for the regions around the local peaks of the integrated intensity. The pointing error was measured to be within 700 in peak to peak by observing CO point sources R Dor or o Cet every 2 hr during this observing term. The spectral intensities were calibrated by employing the standard room-temperature chopper-wheel technique. We observed Ori-KL once a day, and N159W every 2 hr to check the stability of the intensity calibration, and the intensity variation during these observations was less than 13%. We use 0.7 for the main-beam efficiency at 345 GHz, which was measured by TABLE 1 List of Observed GMCs and Transitions GMC a Position a b No. (1) Name (2) Type (3) Region Name (4) 186............ 197............ LMC N J05386904 LMC N J05407008 III III 216............ 153............ 156............ 225............ LMC LMC LMC LMC II III II I 30 Dor N159 N171 N166 N206 N206D GMC 225 N N N N J05446923 J05307106 J05327114 J05477041 (1950) (5) 5 5 5 5 5 5 5 39 40 40 44 31 32 48 0.2 18.2 24.1 52.5 33.9 52.2 35.7 Telescope (1950) (6) Ref. (7) 69 8 0.0 69 47 0.0 70 8 0.0 69 26 39.1 71 10 0.0 71 16 0.0 70 40 0.0 1 1 2 3 4 4 4 12 CO(J ¼ 3Y2) (8) ASTE ASTE ASTE ASTE ASTE ASTE ASTE 12 CO(J ¼ 1Y0) (9) MOPRA MOPRA MOPRA SEST d SEST SEST SEST 13 CO(J ¼ 1Y0) (10) SEST c SEST c ... SEST d SEST SEST SEST Notes.— Units of right ascension are hours, minutes, and seconds, and units of declination are degrees, arcminutes, and arcseconds. Col. (1): Running number of GMC used in Table 1 in Fukui et al. (2007). Col. (2): Name of GMC. Col. (3): Type of GMC. Col. (4): Region name used in this paper. Cols. (5)Y(6): Coordinates used as reference position in each region for these 12CO(J ¼ 3Y2) observations. Col. (7): References for the positions. Cols. (8)Y(10): Telescope used for each observation. a Fukui et al. (2007). b Kawamura et al. (2007). c Johansson et al. (1998). d Garay et al. (2002). References.— (1) Johansson et al. 1998; (2) Kutner et al. 1997; (3) Garay et al. 2002; (4) Mizuno et al. 2001. observing Jupiter. The system noise temperature was typically 300 K in double-sideband (DSB) including the atmosphere toward the zenith. The typical rms noise fluctuations were 0.25 K at a velocity resolution of 0.44 km s1 for a 1 minute integration for an on-position. In total, about 1400 points were observed in equatorial coordinates (B1950). Velocities were relative to the local standard of rest ( LSR). These observations were made remotely from an ASTE operations room in San Pedro de Atacama, Chile, using the network observation system N-COSMOS3 developed by NAOJ (Kamazaki et al. 2005). 2.3. 12 CO(J ¼ 1Y0) and 13 CO(J ¼ 1Y0) 2.3.1. Mopra Observations A 20 ; 120 region, the prominent molecular ridge extending from 30 Doradus southward, was mapped in the J ¼ 1Y0 transition of 12 CO at a frequency of 115 GHz with the 22 m ATNF Mopra telescope, in five runs from 2005 May to September. This region contains the 30 Dor, N159, and N171 regions. The newly implemented on-the-fly (OTF) mode was used, in which the telescope takes data continuously while moving across the sky. Spectra were taken at a 600 spacing so that the 3300 Mopra beam would be well oversampled in the scanning direction; the row spacing was 800 , also assuring oversampling. The typical system noise temperature, Tsys , was 500 K in the single sideband (SSB) toward the zenith. The pointing was checked on the SiO maser R Dor every 2 hr; typical pointing error was less than 500 rms. The digital correlator was configured to output 1024 channels across 64 MHz in each of two orthogonal polarizations. The velocity resolution and coverage were 0.16 and 160 km s1, respectively, at 115 GHz. The observing time was about 100 minutes per field (5 0 ; 5 0 ), providing rms noise fluctuations of 0.34 K at a velocity resolution of 0.65 km s1. Initial spectral processing ( baseline fitting and calibration onto a TA scale) was performed using the livedata task in AIPS++, and the spectra were gridded into data cubes using the AIPS++ gridzilla task. During the gridding, a Gaussian smoothing kernel with a FWHM taken at one-half of the beam size was convolved with the data, so the effective resolution of the output cubes was 3600 . The cubes were then rescaled onto a Tmb scale using an ‘‘extended beam’’ efficiency of 0.55 ( Ladd et al. 2005). This takes into account that sources larger than about 8000 in diameter will couple to both the main beam and the inner error beam of the telescope. The cubes were two-dimensional Gaussian smoothed to a 4500 beam, which is the beam size of SEST at 115 GHz. 0 ing with a grid spacing of 2000 . The system noise temperature was typically 230 K (SSB). The typical integration time was 4 minutes for an on-position, providing rms noise fluctuations of 0.04 K at a velocity resolution of 0.2 km s1. The pointing accuracy was 500 rms. We checked this by observing the SiO maser toward R Dor every 2 hr during this observing term. N159W was observed periodically for pointing checks and intensity calibration. We use 0.8 for the main-beam efficiency at 115 GHz by assuming the main-beam temperature Tmb of N159W to be 6.5Y6.9 K to keep consistency with previous publications (Johansson et al. 1994, 1998). Observations toward the N166 region were also made using the SEST 15 m telescope. The details are described separately by Garay et al. (2002). 0 2.3.2. SEST Observations Observations toward the N206, N206D, and GMC 225 regions in the 12 CO(J ¼ 1Y0) line (115 GHz) were made in 2001 August and 2002 February using the SEST 15 m telescope at La Silla, Chile. The HPBW was 4500 at 115 GHz, the front end was the IRAM 115 SIS receiver, and the spectrometer was a high-resolution AOS with 2048 channels. The typical system noise temperature was 550 K (SSB). The velocity resolution and coverage were 0.2 and 216 km s1, respectively, at 115 GHz. We mapped these three regions in position switching with a grid spacing of 4000 or 2000 . The typical integration time was 1 minute for an on-position, providing rms noise fluctuations of 0.18 K at a velocity resolution of 0.2 km s1. Observations toward the N206, N206D, and GMC 225 regions in the 13 CO(J ¼ 1Y0) line (110 GHz) were made in 2002 February and December 2002, also using the SEST 15 m telescope. We mapped peak positions of 12 CO(J ¼ 1Y0) in position switch- 3. RESULTS 12 We first present the CO(J ¼ 3Y2) images of the clouds at 5 pc resolution and compare them empirically with the 12 CO(J ¼ 1Y0) distribution (x 3.1). Next, we define molecular clumps and estimate the physical parameters of each clump (x 3.2). 3.1. Distributions of the 12 CO(J ¼ 3Y2) Emission 12 In Figure 2, typical CO(J ¼ 3Y2) and 12 CO(J ¼ 1Y0) profiles of the 30 Dor, N159, and GMC 225 regions are presented. The upper panels, Figures 2aY2c, show 12 CO(J ¼ 3Y2) profiles. These illustrate the low noise levels of the present data, typically 0.25 K rms at 0.44 km s1 velocity resolution. Among the present observed positions, the 12 CO(J ¼ 3Y2) intensity is strongest at Tmb 12:3 K toward N159W (Fig. 2b). The lower panels, Figures 2dY2f, show 12 CO(J ¼ 1Y0) profiles toward the same positions, where the gray line indicates 12 CO(J ¼ 3Y2) profiles convolved to a 4500 Gaussian beam following the method described in x 4. The 12CO(J ¼ 3Y2) intensities from the 30 Dor and N159 regions are a little stronger than the 12 CO(J ¼ 1Y0) intensity when convolved to the same resolution. Only toward GMC 225 is the 12 CO(J ¼ 3Y2) intensity about 50% weaker than the 12 CO(J ¼ 1Y0) intensity. The peak velocity and line width are nearly the same between the two transitions in these regions. The distributions of the integrated intensities of 12 CO(J ¼ 3Y2) and 12 CO(J ¼ 1Y0) are shown in Figures 3Y9. Detailed descriptions of each region are presented in the following. 3.1.1. 30 Dor ( Fig. 3) Figures 3a and 3b show the distributions of the 12 CO(J ¼ 3Y2) and 12 CO(J ¼ 1Y0) integrated intensities in the 30 Dor region. We see a general trend that the J ¼ 3Y2 distribution shows more details that are not obvious in the J ¼ 1Y 0 distribution owing to the higher angular resolution and possibly due to the more compact distribution of warmer and denser gas in J ¼ 3Y2 than in J ¼ 1Y 0. We see three peaks corresponding to the 12 CO(J ¼ 1Y 0) peaks, 30 Dor-10, 30 Dor-6, and 30 Dor-12, reported in Johansson et al. (1998). One of them in the north, 30 Dor-06, which is singly peaked in J ¼ 1Y 0, appears to be resolved into two peaks with the present beam. 3.1.2. N159 ( Fig. 4 ) Figures 4a and 4b show the 12 CO(J ¼ 3Y2) and 12 CO(J ¼ 1Y0) distributions in N159. We note that N159W shows the strongest intensity among the present clouds, as well as a very compact peak that is not sufficiently resolved with the present beam. Its radius is estimated to be a few parsecs after deconvolution. N159E also shows a compact distribution with a hint of a subpeak, while N159S shows similar distributions in both J ¼ 1Y0 and J ¼ 3Y2. Fig. 2.—12 CO(J ¼ 3Y2) and 12 CO(J ¼ 1Y0) profiles at selected points. Note that the vertical scale is main-beam temperature, Tmb . The velocities (abscissae) are relative to the LSR. (a) 12 CO(J ¼ 3Y2) profile of 30 Dor. (b) 12 CO(J ¼ 3Y2) profile of N159. (c) 12 CO(J ¼ 3Y2) profile of GMC 225. (d ) 12 CO(J ¼ 1Y0) and smoothed 12 CO(J ¼ 3Y2) profiles of 30 Dor. (e) 12 CO(J ¼ 1Y0) and smoothed 12 CO(J ¼ 3Y2) profiles of N159. ( f ) 12 CO(J ¼ 1Y0) and smoothed 12 CO(J ¼ 3Y2) profiles of GMC 225. [See the electronic edition of the Supplement for a color version of this figure.] The east-west elongation of N159S may be due to the scanning effect of the OTF mapping and needs to be confirmed. J ¼ 1Y 0 as indicated in Table 2 and Table 1 of Kutner et al. (1997). 3.1.4. N166 ( Fig. 6 ) 3.1.3. N171 ( Fig. 5 ) Figures 6a and 6b show the 12 CO(J ¼ 3Y2) and 12 CO(J ¼ 1Y0) distributions in N166. The J ¼ 1Y0 data themselves have already been published by Garay et al. (2002). We see four peaks in both J ¼ 3Y2 and J ¼ 1Y0. Of these peaks, three peaks are named Cloud-B, Cloud-C, and Cloud-D, as reported by Garay et al. (2002), although Cloud-C is resolved into two peaks with the present beam and observing grid. 12 12 Figures 5a and 5b show the CO(J ¼ 3Y2) and CO(J ¼ 1Y0) distributions in N171. We note that the J ¼ 3Y2 emission is weaker than the J ¼ 1Y0 emission. There are multiple velocity components, at VLSR ¼ 225, 230, and 240 km s1 in both J ¼ 3Y2 and 3.1.5. N206 (Fig. 7 ) Figures 7a and 7b show the 12 CO(J ¼ 3Y2) and 12 CO(J ¼ 1Y0) distributions in N206. A J ¼ 1Y0 peak appears to be resolved into two subpeaks and a north-south filamentary structure with the higher angular resolution of the J ¼ 3Y2 line. 3.1.6. N206D (Fig. 8 ) Figures 8a and 8b show the 12 CO(J ¼ 3Y2) and 12 CO(J ¼ 1Y0) distributions in N206D. We can see a head-tail structure in J ¼ 3Y2, although it appears more rounded in J ¼ 1Y0. 3.1.7. GMC 225 ( Fig. 9) Figures 9a and 9b show the 12 CO(J ¼ 3Y2) and 12 CO(J ¼ 1Y0) distributions in GMC 225. We note that J ¼ 3Y2 emissions are weaker than those of J ¼ 1Y0. Fig. 3.— (a) Contour map of 12 CO(J ¼ 3Y2) integrated intensity in the 30 Doradus region. The contour levels are 5, 10, 15, 20, 25, 30, 35, 40, and 45 K km s1. Observed points are indicated by dots. (b) Contour map of 12 CO(J ¼ 1Y0) integrated intensity in the 30 Doradus region. The contour levels are 10, 15, 20, and 25 K km s1. 3.2. Properties of the Clumps We identified clumps in the following way in the J ¼ 3Y2 distributions shown in Figures 3Y9: (1) Pick up local peaks using the integrated intensity. (2) Draw a contour at one-half the peak Fig. 4.— (a) Contour map of 12 CO(J ¼ 3Y2) integrated intensity in the N159 region. The contour levels are 5, 10, 20, 30, 40, 50, 60, 70, 80, 90 and 100 K km s1. Observed points are indicated by dots. (b) Contour map of 12 CO(J ¼ 1Y0) integrated intensity in the N159 region. The contour levels are 10, 20, 30, and 40 K km s1. integrated intensity level and identify it as a clump unless it contains other local peaks. (3) When there are other local peaks inside the contour, draw new contours at the 70% level of each integrated intensity peak. Then, identify clumps separately if their contours do not contain another local peak (the boundary is taken at the ‘‘valley’’ between clumps), or else identify a clump by using the highest contour as a clump boundary. (4) If a spectrum has multiple velocity components with a separation of more than 6 km s1, identify these components to be associated with different clumps. As a result, 32 clumps have been identified: five clumps in the 30 Dor region, 10 clumps in the N159 region, six clumps in the N171 region, five clumps in the N166 region, two clumps in the N206 region, one clump in the N206D region, and three clumps in the GMC 225 region. Their line parameters at the peak positions are shown in Table 2, and their physical properties are listed in Table 3. The clump size, line width, and virial mass lie in the range 1.1Y 12.4 pc, 4.0Y12.8 km s1, and 4:6 ; 10 3 to 2:2 ; 10 5 M, respectively. The smallest clump identified by the procedure above is detected with two observing points, and its size is 1.1 pc. There are several other weak emissions ( below 6 ), and they are detected with one observing point. These emissions are not identified as clumps. When we assume that such small clumps (R 1 pc, dV 7 km s1) exist, their virial mass is 9:3 ; 10 3 M, and this corresponds to the detection limit. There are 18 local peaks above the 6 noise level that are not identified as clumps, because the stronger peaks are located near these local peaks. Their virial mass is 6:5 ; 10 4 M if their size and line width are similar to those of identified clumps (R 7 pc, dV 7 km s1). This seems to correspond to the completeness limit. If the clumps are identified with the 70% level of the peak integrated intensity instead of the 50% level, the clump size changes to one-half the original one, although the line width does not change. Their virial mass also changes to about one-half the original one. Histograms of their physical properties are presented in Figure 10. Typical values are 7 pc, 7 km s1, and 6 ; 10 4 M, in size, line width, and virial mass, respectively. Their line widths are larger than those of the GMCs in our galaxy (e.g., Williams & Blitz 1998; Ikeda et al. 1999; Sun et al. 2006). We note a trend that the masses of these clumps are relatively large compared with those of high-mass cloud cores in the Milky Way, which are of the order of 103 M (e.g., Burton et al. 2005; Yonekura et al. 2005). Precise comparisons will be necessary using same tracers and resolution in the future because those Galactic studies above Fig. 5.— (a) Contour map of 12 CO(J ¼ 3Y2) integrated intensity in the N171 region. The contour levels are 8, 12, 16, and 20 K km s1. Observed points are indicated by dots. (b) Contour map of 12 CO(J ¼ 1Y0) integrated intensity in the N171 region. The contour levels are 8, 12, 16, 20, 24, 28, 32, 36, 40, 44, 48, and 52 K km s1. are based on optically thin millimeter dust continuum and C18O emission. The clumps in the LMC are also fairly compact, with sizes of several parsecs or less. Hereafter, the region names and the numbers of clumps are used to identify clumps (e.g., ‘‘30 Dor No. 1’’). 4. DATA ANALYSIS 4.1. Derivation of Line Intensity Ratios The spatial resolution of the present CO data varies depending on the telescope and frequency. We convolved and regridded these data into the same resolution and position using two-dimensional Gaussian smoothing in order to compare them to derive reliable peak intensity ratios of 12 CO(J ¼ 3Y2) to 12 CO(J ¼ 1Y0) (hereafter, R3Y2=1Y0 ). We made Gaussian fits to each of the 12 CO(J ¼ 3Y2) and 12 CO(J ¼ 1Y0) spectra having a single peak in most cases. We derived peak intensities and FWHM line widths through these fittings. The ratios of the two transitions were then derived as the ratio between peak values of Tmb . The distributions of these ratios are shown along with the 12 CO(J ¼ 1Y0) distributions in panels (a) of Figures 11Y17. The youngest stellar clusters SWB0, whose ages are estimated to be less than 10 Myr, are also shown by gray or black circles ( Bica et al. 1996). The averaged R3Y2=1Y0 over each clump (hereafter, R3Y2=1Y0;clump ) was also derived from the averaged 12 CO(J ¼ 3Y2) and 12 CO(J ¼ 1Y 0) spectra over each clump. A summary of the R3Y2=1Y0;clump for each clump is shown in Table 4. The histogram in Figure 18 shows that the R3Y2=1Y0;clump ranges from 0.2 to 1.6. These ratios will be compared with numerical calculations of radiative transfer in the LVG approximation to derive constraints on density and temperature. 4.2. LVG Analysis 4.2.1. Calculations of LVG Model To estimate the physical properties of the molecular gas in the LMC, we performed an LVG analysis (Goldreich & Kwan 1974) of the CO rotational transitions. The LVG radiative transfer code simulates a spherically symmetric cloud of constant density and temperature with a spherically symmetric velocity distribution proportional to the radius and employs a Castor escape probability formalism (Castor 1970). It solves the equations of statistical equilibrium for the fractional population of CO rotational levels TABLE 2 Observed Properties of 12 CO(J ¼ 3Y2) Clumps Peak Properties Position Region (1) No. (2) 30 Dor ........................... 1 2 3 4 5 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 1 2 3 4 5 1 2 1 1 2 3 N159.............................. N171.............................. N166.............................. N206.............................. N206D........................... GMC 225 ...................... (1950) (3) 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 39 38 38 38 38 40 40 40 40 39 40 39 40 40 39 40 40 40 40 40 40 44 44 45 44 45 31 31 32 47 48 48 8.6 54.6 49.0 49.0 54.6 3.7 35.5 47.1 32.7 49.3 47.2 55.1 29.8 41.4 37.7 0.5 0.5 24.1 24.1 6.4 18.2 58.2 46.8 3.9 32.6 9.6 29.8 46.3 58.4 51.3 25.6 53.9 (1950) (4) Tmb ( K) (5) VLSR ( km s1) (6 ) V ( km s1) (7) Integrated Intensity ( K km s1) (8) 69 6 15 69 8 0 69 4 30 69 3 30 69 9 0 69 47 0 69 46 0 69 46 15 69 52 0 69 47 0 69 51 30 69 45 0 69 50 0 69 51 0 69 45 30 70 12 30 70 10 30 70 9 45 70 8 30 70 9 0 70 9 30 69 26 39 69 27 24 69 29 9 69 23 39 69 30 9 71 9 40 71 8 20 71 15 20 70 41 20 70 41 20 70 40 30 5.2 4.3 3.1 2.6 1.2 12.3 7.6 5.4 3.5 3.0 3.6 4.5 0.91 0.87 0.89 2.5 1.3 1.3 2.3 1.7 1.9 4.5 2.9 2.2 1.0 0.82 5.1 2.3 4.5 1.8 1.2 1.8 250.4 246.9 251.9 248.2 247.4 237.3 233.2 234.6 234.4 233.5 235.6 238.3 234.6 235.7 238.6 224.3 227.3 230.9 231.0 240.6 240.7 227.6 229.4 227.9 236.1 231.6 229.0 231.2 224.6 216.5 216.2 216.4 8.4 6.9 5.4 5.0 4.5 8.1 7.0 6.7 9.7 9.5 5.3 3.4 6.9 8.0 7.9 5.2 8.3 9.2 6.7 4.9 5.1 4.5 6.2 5.5 9.2 4.5 4.6 4.6 4.0 3.9 6.2 3.8 49.5 31.7 22.5 21.1 5.3 106.6 55.5 46.0 35.4 29.7 19.8 16.6 10.5 9.3 8.3 15.4 12.7 18.2 16.1 11.5 10.4 22.1 20.3 14.6 10.1 9.6 26.2 13.5 18.9 9.0 8.6 7.9 Other ID (9) 30Dor-10a 30Dor-12a 30Dor-06a N159-W a N159-Ea N159-Sa 30DOR-CENTER-04b 30DOR-CENTER-03b Cloud-Cc Cloud-Dc Cloud-Bc Notes.— Units of right ascension are hours, minutes, and seconds, and units of declination are degrees, arcminutes, and arcseconds. Col. (1): Region. Col. (2): Running number in each region. Cols. (3)Y(4): Positions of observed points with maximum 12 CO(J ¼ 3Y2) integrated intensities within the 12 CO(J ¼ 3Y2) clumps given in equatorial coordinates. Cols. (5) Y(8): Observed properties of the 12 CO(J ¼ 3Y2) spectra obtained at the peak positions of the 12 CO(J ¼ 3Y2) clumps. The peak main-beam temperature Tmb , VLSR , and the FWHM line width V are derived from a single Gaussian fitting and are given in cols. (5), (6 ), and (7), respectively. In col. (8), the 12 CO(J ¼ 3Y2) integrated intensities at the peak positions of the 12 CO(J ¼ 3Y2) clumps are shown. Col. (9): Another identification based on 12 CO(J ¼ 1Y0) observations with SEST. a Johansson et al. (1998). b Kutner et al. (1997). c Garay et al. (2002). Fig. 6.— (a) Contour map of 12 CO(J ¼ 3Y2) integrated intensity in the N166 region. The contour levels are 6, 8, 10, 12, 14, 16, 18, 20, and 22 K km s1. Observed points are indicated by dots. (b) Contour map of 12 CO(J ¼ 1Y0) integrated intensity in the N166 region. The contour levels are 8, 10, 12, 14, 16, 18, 20, and 22 K km s1. Observed points are indicated by dots. at each density and temperature. It includes the lowest 40 rotational levels of the ground vibrational level and uses the Einstein A and H2 impact rate coefficients of Schöier et al. (2005). The present calculations incorporate the lowest 40 rotational levels of CO in the ground vibrational state over a kinetic temperature range of Tkin ¼ 5Y200 K and a density range of n(H2 ) ¼ 10Y106 cm3. We did not include higher energy levels in the present study, which requires including populations in the lower vibrationally excited states. Therefore, the present work does not cover kinetic temperatures above 200 K, which should be dealt with in the future for analyses of higher submillimeter transitions above J ¼ 4Y3. This imposes a limit of Tkin 200 K in the present study, and even higher temperature is not excluded in general below. We performed calculations for three different CO fractional abundances: X (CO) ¼ ½CO/½H2 ¼ 1 ; 106 , 3 ; 106 , and 1 ; 105 , and three different 12 CO/ 13CO abundance ratios of 20, 25, and 30 (Heikkilä et al. 1999). 4.2.2. Results from 12 CO(J ¼ 3Y2) and Data for 32 Clumps 12 Fig. 8.— (a) Contour map of 12 CO(J ¼ 3Y2) integrated intensity in the N206D region. The contour levels are 4, 6, 8, 10, 12, 14, 16, and 18 K km s1. Observed points are indicated by dots. (b) Contour map of 12 CO(J ¼ 1Y0) integrated intensity in the N206D region. The contour levels are 4, 6, 8, 10, 12, 14, 16, 18, 20, 22, 24, and 26 K km s1. Observed points are indicated by dots. illustrates that we obtain the following lower limits for kinetic temperature and density. For the eight clumps with R3Y2=1Y0;clump 1 (30 Dor Nos. 1, 2, 3, 4, N159 Nos. 1, 2, 6, 8), we estimate Tkin > 30 K and n( H2 ) > 10 3 cm3. For the 24 clumps with R3Y2=1Y0;clump < 1 (30 Dor No. 5, N159 Nos. 3, 4, 5, 7, 9, 10, N171 Nos. 1, 2, 3, 4, 5, 6, N166 Nos. 1, 2, 3, 4, 5, N206 Nos. 1, 2, N206D No. 1, GMC 225 Nos. 1, 2, 3), we estimate Tkin greater than several kelvins and n(H2 ) greater than several times 10 2 cm3. In either case, the lower limits tend to increase with R3Y2=1Y0;clump . 4.2.3. Results from 12 CO(J ¼ 3Y2), 12 CO(J ¼ 1Y0), and 13 CO(J ¼ 1Y0) Data for 13 Clumps We can better constrain these physical parameters when CO(J ¼ 1Y0) data are available. Figure 19 shows the general behavior of the loci of constant R3Y2=1Y0 and constant R12=13 [peak intensity ratio of 12 CO(J ¼ 1Y0) to 13 CO(J ¼ 1Y0)] in the densitytemperature plane. It is recognized that the combination of the two will allow us to determine the parameters relatively well, since the two lines are nearly ‘‘orthogonal’’ in the plane, except for densities higher than 10 4 cm3 ( Fig. 19). 13 CO(J ¼ 1Y0) First, we assume that X(CO) is uniform among the clumps and derive density and temperature using the LVG results. Figure 19a Fig. 7.— (a) Contour map of 12 CO(J ¼ 3Y2) integrated intensity in the N206 region. The contour levels are 4, 6, 8, 10, 12, 14, 16, 18, 20, 22, 24, and 26 K km s1. Observed points are indicated by dots. (b) Contour map of 12 CO(J ¼ 1Y0) integrated intensity in the N206 region. The contour levels are 4, 6, 8, 12, 14, 16, and 18 K km s1. Observed points are indicated by dots. Fig. 9.— (a) Contour map of 12 CO(J ¼ 3Y2) integrated intensity in the GMC 225 region. The contour levels are 3, 4, 5, 6, 7, and 8 K km s1. Observed points are indicated by dots. (b) Contour map of 12 CO(J ¼ 1Y0) integrated intensity in the GMC 225 region. The contour levels are 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, and 19 K km s1. Observed points are indicated by dots. TABLE 3 Physical Properties of 12 CO(J ¼ 3Y2) Clumps Clump Properties Region (1) No. (2) Vclump ( km s1) (3) Rdeconv ( pc) (4) Mvir (;10 4 M) (5) 30 Dor ................ 1 2 3 4 5 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 1 2 3 4 5 1 2 1 1 2 3 12.8 6.9 5.8 4.0 4.6 8.7 7.3 7.6 8.5 10.2 5.2 4.1 6.3 7.1 7.7 6.5 6.9 7.2 6.2 6.1 5.6 6.1 6.5 6.3 9.2 9.5 6.6 5.2 4.0 4.2 6.6 4.9 7.2 3.1 3.1 3.7 1.1 4.7 6.9 4.7 11.3 4.7 4.3 2.3 6.3 2.3 2.3 6.6 8.3 7.8 5.9 7.5 5.9 9.9 9.5 10.3 12.4 4.7 7.1 6.3 7.3 8.8 5.5 8.8 22 2.9 2.0 1.1 0.46 6.8 6.9 5.2 15 9.4 2.2 0.73 4.7 2.2 2.7 5.2 7.4 7.7 4.3 5.3 3.5 6.9 7.6 7.7 20 8.1 5.9 3.2 2.2 3.1 4.6 3.9 N159................... N171................... N166................... N206................... N206D................ GMC 225 ........... Notes.— Col. (1): Region. Col. (2): Running number in each region. Col. (3): Line width, Vclump , derived by using a single Gaussian fitting for a spectrum obtained by averaging all the spectra within a single clump. Col. (4): Deconvolved effective radii, Rdeconv , defined as ½R 2nodeconv (HPBW /2) 2 0:5 . The Rnodeconv is effective radii defined as (A/)0:5 , where A is the total cloud surface area. Col. (5): Virial masses of the 12 CO(J ¼ 3Y2) clumps; a density profile of / r1 was assumed; the virial mass is written as Mvir /M ¼ 190½Vclump (km s1 ) 2 Rdeconv (pc) ( MacLaren et al. 1988). Of the 32 clumps, 13CO(J ¼ 1Y0) data are available for 13 clumps, including four clumps with R3Y2=1Y0;clump 1 (30 Dor Nos. 1, 4, N159 Nos. 1, 2) and nine clumps with R3Y2=1Y 0;clump < 1 (N159 No. 4, N166 Nos. 1, 3, 4, N206 Nos. 1, 2, N206D No. 1, GMC 225 Nos. 1, 3). For these 13 clumps, we made a detailed analysis using the R12=13 at peaks of 12 CO(J ¼ 1Y0) and have obtained the best constraints. We summarize the input parameters for the 13 clumps in columns (3)Y(5) of Table 5. All the data refer to the 12 CO(J ¼ 1Y0) beam size, 4500 ; the higher transition data have been Gaussian smoothed as described in x 4.1. Clump-averaged dv/dr were used for the calculations. The R3Y2=1Y0;clump and R12=13 at the peak of 12 CO(J ¼ 1Y0) were used, and the errors of these ratios are both estimated as 20%, which are derived from errors of absolute intensity calibration. These correspond to 27 and 7 noise levels of 12 CO(J ¼ 3Y2) and 12 CO(J ¼ 1Y0), respectively. This indicates that the errors of the intensity ratios are dominated by the error of the absolute intensity calibration. The way clumps are Fig. 10.—Histograms of physical properties of 12 CO(J ¼ 3Y2) clumps: (a) size, (b) line width, and (c) virial mass. Fig. 11.— (a) Color map of R3Y2=1Y0 , and (b) H flux image of the 30 Doradus region. Contours are 12 CO(J ¼ 3Y2) integrated intensity. Contour levels are the same as in Fig. 3a. Thick lines indicate observed area in 12 CO(J ¼ 3Y2); gray circle indicates position of young cluster (<10 Myr; SWB0). [See the electronic edition of the Supplement for a color version of this figure.] Fig. 12.— (a) Color map of R3Y2=1Y0 , and (b) H flux image of the N159 region. Contours are 12 CO(J ¼ 3Y2) integrated intensity. Contour levels are the same as in Fig. 4a. Thick lines indicate observed area of 12 CO(J ¼ 3Y2); black circles indicate positions of young clusters (<10 Myr; SWB0). [See the electronic edition of the Supplement for a color version of this figure.] Fig. 13.— (a) Color map of R3Y2 =1Y0 , and (b) H flux image of the N171 region. Contours are 12 CO(J ¼ 3Y2) integrated intensity. Contour levels are the same as in Fig. 5a. Thick lines indicate observed area of 12 CO(J ¼ 3Y2). [See the electronic edition of the Supplement for a color version of this figure.] Fig. 14.— (a) Color map of R3Y2=1Y0 , and (b) H flux image of the N166 region. Contours are 12 CO(J ¼ 3Y2) integrated intensity. Contour levels are the same as in Fig. 6a. Thick lines indicate observed area of 12 CO(J ¼ 3Y2). [See the electronic edition of the Supplement for a color version of this figure.] Fig. 15.— (a) Color map of R3Y2 =1Y0 , and (b) H flux image of the N206 region. Contours are 12 CO(J ¼ 3Y2) integrated intensity. Contour levels are the same as in Fig. 7a. Thick lines indicate observed area of 12 CO(J ¼ 3Y2). [See the electronic edition of the Supplement for a color version of this figure.] Fig. 16.— (a) Color map of R3Y2 =1Y0 , and (b) H flux image of the N206D region. Contours are 12 CO(J ¼ 3Y2) integrated intensity. Contour levels are the same as in Fig. 8a. Thick lines indicate observed area of 12 CO(J ¼ 3Y2). [See the electronic edition of the Supplement for a color version of this figure.] Clumps that we detected are distributed continuously from cool (10Y30 K) to warm (higher than 30Y200 K), and warm clumps are distributed from less dense (10 3 cm3 ) to dense (10 3.5 Y 10 5 cm3 ), although cool clumps are all less dense. The three cases shown in Figure 20 represent typical cases, ‘‘warm and dense’’ [Tkin k 30Y200 K, n(H2 ) 10 3:5 Y10 5 cm3 ], ‘‘warm and less dense’’ [Tkin k 30Y200 K, n(H2 ) 10 3 cm3 ], and ‘‘cool and less dense’’ [Tkin 10Y30 K, n(H2 ) 10 3 cm3 ]. 4.2.4. Effects of X (CO) Fig. 17.— (a) Color map of R3Y2=1Y0 , and (b) H flux image of the GMC 225 region. Contours are 12 CO(J ¼ 3Y2) integrated intensity. Contour levels are the same as in Fig. 9a. Thick lines indicate observed area of 12 CO(J ¼ 3Y2). [See the electronic edition of the Supplement for a color version of this figure.] defined does not change R3Y2=1Y0;clump but changes dv/dr, and dv/dr is about 2 times as large as the original one, when the clumps are identified with the 70% level of the peak integrated intensity level. This, however, does not affect the results of the LVG calculations. We describe three typical cases: (a) 30 Dor No. 1, (b) N206 No. 1, and (c) GMC 225 No. 1 are shown in Figure 20. A fractional CO abundance of X(CO) of 3 ; 106 was used throughout. The horizontal axis is molecular hydrogen density [n(H2)], and the vertical axis is the gas kinetic temperature (Tkin ). Solid lines represent R3Y2=1Y0;clump , and dashed lines represent R12=13 at 12 CO(J ¼ 1Y0) peaks. Hatched areas indicate the overlap regions of these two ratios within the errors, which is allowed from the observed ratios. It also includes the uncertainty due to a possible variation of the 12 CO/ 13CO abundance ratio from 20 to 30. Hereafter, we call clumps above 30 K ‘‘warm’’ and those below ‘‘cold.’’ Clumps with densities greater than 10 3.5 cm3 are referred to as ‘‘dense’’ and those at lower densities as ‘‘less dense.’’ 30 Dor No. 1 exhibits a R3Y2=1Y0;clump of 1.4, suggesting that it is a warm (Tkin ¼ 50Y200 K) and dense (n ¼ 10 3 Y10 5 cm3 ) clump (Fig. 20a). N206 No. 1 ( Fig. 20b) and GMC 225 No. 1 (Fig. 20c) show R3Y2=1Y0;clump 0:7 and 0.4, respectively. Figures 20b and 20c indicate that they are warm (Tkin ¼ 30Y200 K) and less dense clumps (n ¼ 10 3 cm3) and cold (Tkin ¼ 15Y40 K) and less dense (n ¼ 10 3 cm3) clumps, respectively. We find that R12=13 is useful to discriminate the temperature difference except in the case of 30 Dor No. 1. All the figures for the 13 clumps for all different fractional CO abundances (1 ; 106 , 3 ; 106 , and 1 ; 105 ) are given in Appendix A. In Table 5 we present the results for the 13 clumps for a fixed fractional CO abundance of X (CO) ¼ 3 ; 106 . The results of previous studies are also summarized in Table 5 for comparison. We find that the present results are the most extensive among these studies in terms of the number of samples, while they are basically consistent with previous studies for the individual clumps or clouds. There are relatively large differences in both density and temperature in N159 No. 4 (N159S) only, and this is because of the high-density tracers used in Heikkilä et al. (1999). Figure 21 summarizes the calculated densities and temperatures for the 13 clumps. We see first that the temperature ranges from 10 K up to more than 200 K and density from 10 3 to 10 5 cm3. We now discuss the possible effect of changing X(CO), as summarized in Table 6. If we adopt X (CO) ¼ 1 ; 106 , the contours of R3Y2=1Y0 shift to lower temperature and the contours of R12=13 shift to higher density. Accordingly, the solution shifts to lower temperature, higher density. If we adopt X (CO) ¼ 1 ; 105 , the R3Y2=1Y0 contours shift to higher temperature and the R12=13 contours shift to lower density. Accordingly, the solution shifts to higher temperature, lower density (Table 6). Next, we vary X(CO) from clump to clump. Heikkilä et al. (1999) estimate X(CO) in 30 Dor No. 1 (30 Dor-10), N159 No. 1 ( N159W), and N159 No. 4 ( N159S) to be 1 ; 106 , 1 ; 105 , and 3 ; 106 , respectively. Table 6 indicates that 30 Dor No. 1 shows similar values to Heikkilä et al. (1999) whereas N159 No. 1 and No. 4 become warmer and lower in density than in Heikkilä et al. (1999). This discrepancy may be due to the fact that the high-density tracers used by Heikkilä et al. (1999) are not used in the present study. To summarize, the assumption of uniform fractional abundance, X(CO), is fairly good and the present results do not show significant difference even if we adopt the different fractional abundances used by previous authors. 4.3. Comparisons to H Flux 4.3.1. Relation between R3Y2=1Y0 and H Flux We converted the H data (Kim et al. 1999) toward the present clouds using the method given in Appendix B. The typical background level of the H flux is 1012 ergs s1 cm2 at the 4000 scale and ranges up to 1010 ergs s1 cm2 toward strong H ii regions (Fig. 22). These data were regridded into the 12 CO(J ¼ 3Y2) data grids. H flux images with the youngest stellar clusters SWB0 (Bica et al. 1996; younger than 10 Myr) are shown in panels (b) of Figures 11Y17, and the 12 CO(J ¼ 3Y2) contours from Figures 3Y9 are overlaid for comparison in these figures. Figures 11Y17 indicate a clear trend that the R3Y2=1Y0 is enhanced to 1.0Y1.5 toward H ii regions or clouds with young clusters, as in the 30 Dor and N159 regions. On the other hand, the ratio is low, less than 1.0, toward clumps with neither H ii regions nor clusters, as in GMC 225. We also note that the ratio is enhanced toward the regions where H is intense or toward the interfaces between clouds and H ii regions. A summary of the averaged H flux over each clump is shown in Table 4. It is not fully guaranteed that the H emission is actually in contact with the molecular gas, and some of the apparent coincidence could be fortuitous. Nonetheless, previous studies comparing CO and H flux indicate a strong correlation between them and lend support to the assumption that almost all the coincidences indicate an actual physical association (Fukui et al. 1999; Yamaguchi et al. 2001; A. Kawamura et al. 2008, in preparation). Figure 23 shows the correlation between R3Y2=1Y0;clump and the averaged H flux over each clump. It is clear that R3Y2=1Y0;clump is well correlated with the averaged H flux, with a correlation coefficient of 0.79. This is a fairly good empirical relationship and should be tested to see if it holds true in other galaxies. The 12 TABLE 4 CO(J ¼ 3Y2)/ 12 CO(J ¼ 1Y0) Ratio and H Flux Region (1) No. (2) Tmb (3Y2) ( K) (3) Tmb (1Y0) ( K) (4) R3Y2=1Y0;clump (5) Averaged H Flux (;1012 ergs s1 cm2 ) (6 ) 30 Dor .............................. 1 2 3 4 5 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 1 2 3 4 5 1 2 1 1 2 3 1.63 1.76 1.56 1.98 0.73 5.50 3.69 2.92 2.33 1.84 1.61 1.67 0.78 0.80 0.57 1.04 0.69 1.02 1.34 0.88 0.87 1.97 1.59 1.17 0.79 0.62 1.83 1.34 2.26 1.04 0.55 0.77 1.20 1.12 1.13 1.58 0.87 5.04 3.34 3.25 3.34 2.19 1.57 1.98 0.57 0.99 0.84 2.20 1.55 3.09 3.89 1.92 2.72 3.09 2.60 1.92 1.26 0.92 2.39 2.44 4.70 2.81 1.62 2.09 1.4 1.6 1.2 1.3 0.8 1.1 1.1 0.9 0.7 0.8 1.0 0.9 1.4 0.8 0.7 0.5 0.4 0.3 0.3 0.4 0.3 0.6 0.6 0.6 0.6 0.7 0.8 0.5 0.5 0.4 0.3 0.4 56.49 73.37 24.09 12.83 39.71 9.28 9.37 6.36 1.14 1.93 1.42 3.14 1.52 1.26 1.42 1.00 0.93 0.97 0.98 0.95 0.97 1.05 1.01 1.06 1.31 1.06 3.26 2.68 1.11 0.93 0.91 0.92 N159................................. N171................................. N166................................. N206................................. N206D.............................. GMC 225 ......................... Notes.— Col. (1): Region. Col. (2): Running number in each region. Col. (3): The peak main-beam temperature, Tmb , of the CO(J ¼ 3Y2) spectra derived by using a single Gaussian fitting for a spectrum obtained by averaging all the spectra to the beam within a single clump. The intensities refer to the 12 CO(J ¼ 1Y0) beam size (4500 ). Col. (4): Peak main-beam temperature, Tmb , of the 12 CO(J ¼ 1Y0) spectra derived by using a single Gaussian fitting for a spectrum obtained by averaging all the spectra within a single clump. Col. (5 ): Ratios of Tmb (3Y2) to Tmb (1Y0). Col. (6 ): H flux obtained by averaging within a single clump. 12 clumps in type III and II GMCs. The effects of H emission or nearby clusters on the molecular gas are perhaps local phenomena as indicated by the comparison in Figures 11Y17 for the individual clumps. 5. DISCUSSIONS 5.1. Dense and Compact Clumps as Candidates for Protocluster Condensations Fig. 18.— Histogram of clump-averaged peak intensity ratio of to 12 CO(J ¼ 1Y0) (R3Y2=1Y0;clump ). 12 CO(J ¼ 3Y2) relation suggests that higher R3Y2=1Y0 reflect higher temperatures or higher densities. It is notable that the clumps with averaged H flux greater than 1011 ergs s1 cm2 correspond to high R3Y2=1Y0;clump of 1.0Y1.5. In the barred spiral galaxy M83, the CO(J ¼ 3Y2)/(J ¼ 1Y0) integrated intensity ratio exceeds unity at the nucleus, whereas the ratio gradually decreases to 0.6Y0.7 with distance from the center. The ratio is constant through the disk region (Muraoka et al. 2007). We have carried out submillimeter 12 CO(J ¼ 3Y2) observations of GMCs in the LMC that are most extensive and highly sensitive compared to the previous studies. Six GMCs were selected based on the NANTEN CO survey of the LMC, including three type III GMCs actively forming O stars, in addition to three type I/II GMCs that are quiet in O star formation or cluster formation, although the formation of low- to intermediate-mass stars is not excluded. The spatial resolution of 5 pc and the high sensitivity allowed us to identify 32 molecular clumps in these GMCs and to reveal significant details of the warm and dense molecular gas with n(H2 ) 10 3 Y105 cm3 and Tkin 10Y200 K. The typical mass of the molecular clumps is large, in the range of 5 ; 10 3 to 2 ; 105 M with radii of 1Y12 pc. Of all of our objects, N159 No. 1 or W shows the strongest concentration of mass of 7 ; 10 4 M within a radius of 5 pc. The masses seem to be larger than those of typical Milky Way GMCs such as those in the Car region (e.g., Yonekura et al. 2005), although the properties of these Galactic GMCs are based on optically thin C18O data. We suggest that these are good candidates for the precursors of rich superclusters like R136 in 30 Dor, which includes more than 10 4 stars in a small volume with a radius of 1 pc. It is of particular interest to look for even denser gas toward them in higher excitation transitions of the submillimeter region. 5.2. Density and Temperature of the Clumps and Implications 4.3.2. Comparison with Physical Properties and H Emission Figure 24 shows plots of density and temperature as functions of averaged H flux and gives us another insight into these properties. Where averaged H is strong (1011 ergs s1 cm2) the clumps are always warm, at around Tkin ¼ 100 K or more. On the other hand, when averaged H is weak, density is always low but temperature can be either high or low. The high-density clumps show high R3Y2=1Y0;clump of 1.0Y1.5 and are associated with strong H flux, while the low-density clumps show low R3Y2=1Y0;clump of 0.5Y1.0 with weak H flux. Since we are averaging the H intensity in each clump, we may be diluting the localized low H flux toward some of the The results of our LVG analysis indicate that clumps are distributed from cool to warm in temperature and from less dense to dense in density. These differences of clump properties in density and temperature show good correspondence with the GMC types based on the star formation activity, as well as with the H emission of ionized gas associated with each clump. Clumps in type III GMCs are warm (Tkin 30Y200 K) and are either dense [n(H2 ) 10 3:5 Y105 cm3] or less dense [n(H2 ) 10 3 cm3]. Clumps in type II GMCs are either warm (Tkin 30Y200 K) or cool (Tkin 10Y30 K) and less dense [n(H2 ) 10 3 cm3]. Clumps in type I GMCs are cool (Tkin 10Y30 K) and less dense [n(H2 ) 10 3 cm3 ]. The physical parameters of clumps are Fig. 19.— Contour plots of LVG analysis for reference. Contours are (a) R3Y2=1Y0 , (b) R12=13 , and (c) a þ b. Here X (CO) ¼ 3 ; 106 , dv/dr ¼ 1:0 km s1 pc1, and the abundance ratio of 12 CO/ 13CO is 25. TABLE 5 Summary of LVG and Mean Escape Probability Analyses J98 This Work H99 B05 SC B05 CDC B05 HTC K06 R12=13 Region (1) No. (2) dv/dr ( km s1 pc1) (3) R3Y2=1Y0;clump (4) Value (5) Ref. (6) 30 Dor ............... 1 4 1 2 4 1 3 4 1 2 1 1 3 0.9 0.5 0.9 0.5 0.4 0.3 0.3 0.4 0.5 0.4 0.3 0.2 0.3 1.4 1.3 1.1 1.1 0.7 0.6 0.6 0.6 0.8 0.5 0.5 0.4 0.4 11.5 17.7 8.6 11.6 8.5 10.5 12.6 16.2 14.1 9.8 4.8 6.6 6.7 J98 J98 J98 J98 J98 G02 G02 G02 This work This work This work This work This work N159.................. N166.................. N206.................. N206D............... GMC 225 .......... n( H2) (cm3) (7) 3 1 3 1 1 5 3 4 5 6 1 5 7 ; ; ; ; ; ; ; ; ; ; ; ; ; 10 3 10 3 10 3 10 3 10 3 10 2 10 2 10 2 10 2 10 2 10 3 10 2 10 2 to to to to to to to to to to to to to 3 1 8 3 6 2 2 2 3 2 3 2 2 Tkin (8) ; ; ; ; ; ; ; ; ; ; ; ; ; 10 5 10 5 10 5 10 5 10 3 10 3 10 3 10 3 10 3 10 3 10 3 10 3 10 3 >50 >60 >30 >40 20 Y60 25 Y150 >30 >40 >35 20 Y80 10 Y20 15 Y20 10 Y40 n( H2) (cm3) (9) Tkin (10) 104.5 104.5 104.5 40Y80 >20 16Y23 n( H2) (cm3) (11) Tkin (12) 1 ; 10 5 50 3 ; 10 5 25 1 ; 10 5 10 n( H2) (cm3) (13) 10 3 Tkin (14) 150 n( H2) (cm3) (15) Tkin (16) n( H2) (cm3) (17) Tkin (18) 10 5 20 10 2 100 10 4 20 10 2 to 10 3 30Y 60 n( H2) (cm3) (19) Tkin (20) 10 4.3 100 Notes.— Cols. (7) and (8): Results of LVG analysis. Cols. (9)Y(20): Results of previous studies, including both LVG and mean escape probability analyses. References.— (J98) Johansson et al. 1998; (G02) Garay et al. 2002; (H99) Heikkilä et al. 1999; ( B05 SC) ‘‘Single-component fit’’ in Bolatto et al. 2005; ( B05 CDC, B05 HTC) ‘‘Cold Dense Component’’ and ‘‘Hot Tenuous Component,’’ respectively, of the ‘‘Dual-component fit’’ in Bolatto et al. 2005; (K06) Kim 2006. Fig. 20.— Contour plots of LVG analysis of three clumps: (a) 30 Dor No. 1, (b) N206 No. 1, and (c) GMC 225 No. 1. The vertical axis is kinetic temperature Tkin , and the horizontal axis is molecular hydrogen density n( H2). Solid lines indicate R3Y2=1Y0;clump , and dashed lines indicate R12=13 . Hatched areas are the regions in which these two ratios overlap within intensity calibration errors of 20% and uncertainty due to a possible variation of 12 CO/ 13CO abundance ratio from 20 to 30. generally correlated with the star formation activity of GMCs and can perhaps be interpreted in terms of evolutionary effects. Our interpretation is that differences of clump density and temperature represent an evolutionary sequence of GMCs in terms of density increase leading to star formation; type I /II GMCs are at a young phase of star formation where density has not yet reached high enough values to cause active massive star formation, and type III GMCs represent the later phase where the average density is higher, including both high- and low-density subtypes. The high-density clumps in type III GMCs show high R3Y2=1Y0;clump of 1.0Y1.5 and are associated with strong H flux, while the lowdensity clumps in type III GMCs show low R3Y2=1Y0;clump of 0.5 Y 1.0 with weak H flux. We suggest two alternative ideas to explain the density difference of the clumps in type III GMCs; one is that density is being enhanced by shock compression driven by H ii regions, and the other is that gravitational condensation of each clump plays a role in the density increase. The former may be difficult because the shock front may occupy a small volume that is likely missed with the present 5 pc beam. It seems thus favorable that the latter scenario is working mainly to enhance density. The present study, which resolved the smaller clumps in GMCs at 5Y10 pc scales, indicates that the clumps may have physical properties affected by local properties such as the H distribution. It should be interesting to investigate the variations among these internal clumps and their relation to star formation. 5.3. FUV Heating of the Molecular Gas in the LMC The present findings that the R3Y2=1Y0;clump is well correlated with H flux suggest that the heating of molecular gas by far- Fig. 21.— Plot of LVG results. The vertical axis is kinetic temperature, Tkin , and the horizontal axis is molecular hydrogen density n( H2). [See the electronic edition of the Supplement for a color version of this figure.] ultraviolet ( FUV ) photons may be effective in the LMC where the dust opacity is lower and the FUV intensity is higher than in the Milky Way. The molecular gas in the Milky Way is mainly heated by cosmic-ray protons of 100 MeV as discussed by a number of authors, although the surface layers of molecular clouds with small visual extinctions at AV around a few magnitudes or less may be dominated by the FUV heating (e.g., Kaufman et al. 1999). Some authors have made detailed calculations of gas heating and cooling under the effects of FUV radiation fields (Kaufman et al. 1999). We try to present a picture that can be applied to the present results below. First, the gas temperature is determined through the balance between the cooling and heating. According to Table 4 in Goldsmith & Langer (1978) the total cooling rate is 6:8 ; 1027 T 2:2 ergs cm3 s1 for X (CO)/(dv/dr) ¼ 4 ; 105 and n(H2 ) ¼ 10 3 cm3. In the 30 Dor region, since X (CO)/(dv/dr) ¼ 3 ; 106/0:8 ¼ 3:75 ; 106 and this value is 10 times lower than the value used in Goldsmith & Langer (1978), n( H2 ) ¼10 4 cm3 can be read 103 cm3; then the cooling rate is estimated as 1:7 ; 1022 ergs cm3 s1 for T ¼ 100 K. This value is a factor 2Y3 smaller than that of Galactic clouds with n(H2 ) ¼ 10 4 cm3 and T ¼ 50 K. We assume that the heating by cosmic-ray electrons is not important in the LMC. This assumption is not directly confirmed, but it is consistent with the low nonthermal fraction of the LMC’s radio continuum emission (e.g., Hughes et al. 2006) and studies that suggest that a significant fraction of cosmic-ray electrons are able to escape from low-luminosity galaxies (e.g., Bell 2003; Skillman & Klein 1988). The FUV flux (G0 ) is estimated as 3500 for 30 Doradus (Bolatto et al. 1999; Poglitsch et al. 1995; Werner et al. 1978; Israel & Koornneef 1979) and 300 for N159 ( Bolatto et al. 1999; Israel et al. 1996; Israel & Koornneef 1979). In Orion, it is estimated as 25 (Bolatto et al. 1999; Stacey et al. 1993). The FUV flux in the LMC is larger than that in the Milky Way. Photodissociation region ( PDR) models are calculated by Kaufman et al. (1999) that incorporate the chemical and physical processes that form and destruct atoms or molecules, as well as ionization effects. Figure 1 of Kaufman et al. (1999) shows the kinetic temperature for a molecular gas layer with density of n (cm3) under FUV flux of G0 at the surface. PDR surface temperatures are estimated as listed in Table 7. These indicate that temperature becomes as high as 100Y300 K on the PDR surface under the conditions of the clumps in type III GMCs in the LMC. These temperatures are basically consistent with the temperatures of the warm clumps in the present sample. TABLE 6 Effect of X(CO) X(CO) Conditions Parameter 30 Dor No. 1 N159 No. 1 N159 No. 4 3 ; 106 3 ; 103 to 8 ; 105 >30 1 ; 105 2 ; 103 to 2 ; 105 >45 3 ; 106 1 ; 103 to 6 ; 103 20 Y60 3 ; 106 1 ; 103 to 6 ; 103 20 Y60 1.2 ; 105 3 ; 105 25 4.6 ; 106 1 ; 105 10 This Work Uniform X(CO) ................. Different X(CO) ................ X(CO) n( H2 ) (cm3 ) Tkin ( K) X(CO) n( H2 ) (cm3 ) Tkin ( K) 3 ; 106 3 ; 10 3 to 3 ; 105 >50 1 ; 106 6 ; 10 3 to 1 ; 106 >35 Heikkilä et al. (1999) Different X(CO) ................ X(CO) n( H2 ) (cm3 ) Tkin ( K) 1.4 ; 106 1 ; 105 50 Generally speaking, at a scale of 10 pc, Tkin 100 K seems to be higher than the kinetic temperatures typical in Milky Way GMCs, where the Milky Way values are usually derived from the 12 CO(J ¼ 1Y0) emission only (e.g., Car Tkin 50 K by Yonekura et al. 2005). This suggests that the heating of molecular clouds may be stronger in the LMC than in the Milky Way and the molecular temperature may be higher. If this is correct, the lower metallicity, resulting in lower extinction, is the basic cause for the higher temperature in addition to the stronger FUV field in the LMC. We note in the end that this higher temperature in the molecular gas possibly leads to an increase of the Jeans mass of molecular clumps, which may favor the formation of rich superclusters in the LMC. This is consistent with the higher mass of the molecular clumps, which may represent precursors of the clusters. The present work has undertaken to sample six GMCs (seven regions) to have a uniform determination of the density and temperature in the LMC. The number of GMCs is still limited to 6 among 300 detected with NANTEN. We should make more efforts to collect appropriate data sets in the submillimeter wavelengths to improve our understanding of the cloud properties. Fig. 22.— Histogram of clump-averaged H flux. Background level is nearly 1012 ergs s1 cm2. NANTEN2, ASTE, and others will certainly be powerful in achieving this goal. 6. SUMMARY We summarize the results as follows. 1. We have used the ASTE 10 m telescope to obtain the distribution of 12 CO(J ¼ 3Y2) emission at 345 GHz toward six GMCs (seven regions) in the LMC. We have identified 32 clumps in these GMCs at 5 pc resolution. The radius, line width, and virial mass are estimated as 1.1Y12.4 pc (7 pc), 4.0Y12.8 km s1 (7 km s1), and 4:6 ; 10 3 to 2:2 ; 10 5 M (6 ; 10 4 M), respectively, with the average values in the parenthesis. 2. We have compared the present results with LVG radiative line transfer calculations in order to obtain the density and temperature estimated for the 13 clumps using R3Y2=1Y0;clump and R12=13 at 4500 resolution. The clumps are distributed from cool (10 Y 30 K) to warm (more than 30Y200 K) and from less dense [n(H2 ) 10 3 cm3 ] to dense [n( H2 ) 10 3:5 Y105 cm3]. 3. The H flux toward these clumps is well correlated with the 12 CO(J ¼ 3Y2)/12 CO(J ¼ 1Y0) ratio, R3Y2=1Y0;clump , and clumps with H fluxes greater than 1011 ergs s1 cm2 have large R3Y2=1Y0;clump of 1.5. The 12 CO(J ¼ 1Y0) data were taken with the SEST 15 m and Mopra 22 m telescopes. 4. The typical mass of the molecular clumps ranges from 5 ; 10 3 to 2 ; 10 5 M with radii of 1Y12 pc. Of all of our objects, N159 No. 1 or W shows the strongest concentration of mass of 7 ; 104 M within a radius of 5 pc. We suggest that these are good candidates for the precursors of rich superclusters like R136 in 30 Dor. Fig. 23.— Plots of R3Y2=1Y0;clump as a function of clump-averaged H flux by (a) region and (b) GMC type. [See the electronic edition of the Supplement for a color version of this figure.] TABLE 7 Estimated PDR Surface Temperatures Region (1) n (cm3) (2) 30 Dor ....................... N159.......................... 10 4 10 4 G0 (3) Ts a ( K) (4) References (5) 3500 300 300 100 1, 2, 3, 4 1, 4, 5 Notes.— Col. (1): Regions. Col. (2): Gas density. Col. (3): FUV flux in units of local interstellar value; 1:6 ; 103 ergs cm2 s1. Col. (4): Derived PDR surface temperature of the atomic gas. Col. (5): References of FUV flux. a Estimation is done by using Fig. 1 of Kaufman et al. 1999. References.— (1) Bolatto et al. 1999; (2) Poglitsch et al. 1995; (3) Werner et al. 1978; (4) Israel & Koornneef 1979; (5) Israel et al. 1996. more, the high temperature in the molecular gas possibly leads to an increase of the Jeans mass of molecular clumps, which may favor the formation of rich superclusters. Fig. 24.— Plots of physical properties as a function of clump-averaged H flux. (a) n( H2 ) by region. (b) n( H2 ) by GMC type. (c) Tkin by region. (d ) Tkin by GMC type. [See the electronic edition of the Supplement for a color version of this figure.] 5. We suggest that differences of clump properties represent an evolutionary sequence of GMCs in terms of density increase leading to star formation. Type I /II GMCs are at a young phase of star formation where density has not yet reached high enough values to cause active massive star formation, and type III GMCs represent the later phase where the average density is higher, including both high- and low-density subtypes. 6. The high kinetic temperature correlated with H flux suggests that FUV heating is dominant in the molecular gas of the LMC. The low fraction of nonthermal radio continuum emission and calculations of PDR models support this suggestion. Further- A part of this study was financially supported by MEXT Grantsin-Aid for Scientific Research on Priority Area (15071202 and 15071203) and by JSPS ( No. 14102003, core-to-core program 17004, and No. 18684003). T. M. is supported by JSPS Research Fellowships for Young Scientists. M. R. and S. N. are supported by the Chilean Center for Astrophysics FONDAP 15010003. S. K. was supported in part by the Korea Science and Engineering Foundation (KOSEF) under a cooperative agreement with the Astrophysical Research Center of the Structure and Evolution of the Cosmos (ARCSEC). ASTE is a joint project between Japan and Chile. The telescope is operated by the ASTE team, including NAOJ, University of Tokyo, Nagoya University, Osaka Prefecture University, and Universidad de Chile. We are grateful to all the members of the ASTE team. APPENDIX A COMPLETE LVG RESULTS OF ALL CLUMPS Figures 25Y37 show the complete LVG results described in x 4.2. We present all cases: X(CO) of (a) 1 ; 106 , (b) 3 ; 106 , and (c) 1 ; 105 for 13 clumps. Fig. 25.— Contour plots of LVG analysis of 30 Dor No. 1. The vertical axis is kinetic temperature, Tkin , and the horizontal axis is molecular hydrogen density, n( H2). Solid lines represent R3Y2=1Y0;clump , and dashed lines represent R12=13 . Hatched areas are the regions in which these two ratios overlap within intensity calibration errors and uncertainty due to a possible variation of 12 CO/ 13CO abundance ratio. Here X(CO) = (a) 1 ; 106 , (b) 3 ; 106 , and (c) 1 ; 105 . Fig. 26.— Same as Fig. 25, but for 30 Dor No. 4. Fig. 27.— Same as Fig. 25, but for N159 No. 1. Fig. 28.— Same as Fig. 25, but for N159 No. 2. Fig. 29.— Same as Fig. 25, but for N159 No. 4. Fig. 30.— Same as Fig. 25, but for N166 No. 1. Fig. 31.—Same as Fig. 25, but for N166 No. 3. Fig. 32.— Same as Fig. 25, but for N166 No. 4. Fig. 33.—Same as Fig. 25, but for N206 No. 1. Fig. 34.— Same as Fig. 25, but for N206 No. 2. Fig. 35.— Same as Fig. 25, but for N206D No. 1. Fig. 36.— Same as Fig. 25, but for GMC 225 No. 1. Fig. 37.— Same as Fig. 25, but for GMC 225 No. 3. Fig. 38.— Plot of relation between integrated H flux and FITS value. The solid line represents the least-squares fit. Fig. 39.— Plots of physical properties as a function of R3Y2=1Y0;clump . (a) n(H2) by region. (b) n( H2) by GMC type. (c) Tkin by region. (d ) Tkin by GMC type. [See the electronic edition of the Supplement for a color version of this figure.] Fig. 40.— Same as Fig. 39, but as a function of R12=13 . [See the electronic edition of the Supplement for a color version of this figure.] APPENDIX B H FLUX We here describe the method of scaling data values in the FITS cube (H image in Kim et al. [1999], hereafter ‘‘Kim’s FITS’’). The data values in Kim’s FITS are not flux scale, and the calibration is needed for quantitative comparison with CO clouds and R3Y2=1Y0 . The procedure is as follows. (1) Sum up the data values of Kim’s FITS inside apertures, which are listed in Kennicutt & Hodge (1986) for each listed H ii region. (2) Plot cataloged values of H flux in Kennicutt & Hodge (1986) as a function of summed values derived in step 1. They are well fitted by a power function of y ¼ 5:16 ; 1015 x0:9 (c:c: ¼ 0:94; Fig. 38). (3) Convert data values of Kim’s FITS to H flux scale (ergs s1 cm2) using the function derived above. APPENDIX C LVG RESULTS IN THE OTHER PLANES C1. PHYSICAL PROPERTIES: R3Y2=1Y0;clump (FIG. 39) The density plots (Figs. 39a and 39b) show that higher R3Y2=1Y0;clump (>1.0) correspond to higher densities of 10 3 Y10 5 cm3, while lower R3Y2=1Y0;clump (<1.0) correspond to lower densities of around 10 3 cm3. The temperature plots (Figs. 39c and 39d ) show that higher R3Y2=1Y0;clump (> 0.5) correspond to higher temperatures of >30 K, while lower R3Y2=1Y0;clump (<0.5) correspond to lower temperatures of <30 K. Then, roughly speaking, we can say that clumps with R3Y2=1Y0;clump lower than 0.5 have lower densities of around 10 3 cm3 and lower temperatures of <30 K, clumps with R3Y2=1Y0;clump of 0.5Y1.0 have lower density around 10 3 cm3 and higher temperatures of >30 K, and clumps with R3Y2=1Y0;clump higher than 1.0 have higher densities of 10 3 Y10 5 cm3 and higher temperatures of >30 K, although ratios, densities, and temperatures are with large error bars. Of course, there are great benefits to using R12=13 in LVG analyses. We could not have obtained the above results with only R3Y2=1Y0, as mentioned in x 4. Fig. 41.— Plots of R12=13 as a function of clump-averaged H flux (a) by region and (b) by GMC type. [See the electronic edition of the Supplement for a color version of this figure.] The density ( Figs. 40a and 40b) does not show a significant correlation with R12=13 . The temperature (Figs. 40c and 40d ) shows a good correlation with R12=13 ; that is, higher ratios indicate higher temperatures. Usually, R12=13 correspond to density, but in this case, due to the LVG analysis using both R3Y2=1Y0 and R12=13 , larger R12=13 indicate lower density and lower density tends to higher temperature. C3. R12=13 -H FLUX ( FIG. 41) There is no significant relation between R12=13 and H flux, in contrast to the case for the R3Y2=1Y0 ratio. REFERENCES Bell, E. 2003, ApJ, 586, 794 Kamazaki, T., et al. 2005, in ASP Conf. Ser. 347, Astronomical Data Analysis Bica, E., Clariá, J. J., Dottori, H., Santos, J. F., Jr., & Piatti, A. E. 1996, ApJS, Software and Systems XIV, ed. P. Shopbell, M. Britton, & R. Ebert (San 102, 57 Francisco: ASP), 533 Blitz, L., Fukui, Y., Kawamura, A., Leroy, A., Mizuno, N., & Rosolowsky, E. Kaufman, M. J., Wolfire, M. G., Hollenbach, D. J., & Luhman, M. L. 1999, 2007, in Protostar and Planets V, ed. B. Reipurth, D. Jewitt, & K. Keil ApJ, 527, 795 ( Tucson: Univ. Arizona Press), 81 Kawamura, A., Minamidani, T., Mizuno, Y., Onishi, T., Mizuno, N., Mizuno, Bolatto, A. D., Israel, F. P., & Martin, C. L. 2005, ApJ, 633, 210 A., & Fukui, Y. 2007, in IAU Symp. 237, Triggered Star Formation in a Bolatto, A. D., Jackson, J. M., & Ingalls, J. G. 1999, ApJ, 513, 275 Turbulent ISM, ed. B. G. Elmegreen & J. Palous (Cambridge: Cambridge Burton, M. G., Hill, T., Longmore, N., Purcell, C. R., & Walsh, A. J. 2005, in Univ. Press), 101 IAU Symp. 227, Massive Star Birth: A Crossroads of Astrophysics, ed. Kennicutt, R. C., Jr., & Hodge, P. W. 1986, ApJ, 306, 130 R. Cesaroni et al. (Cambridge: Cambridge Univ. Press), 157 Kim, S. 2006, PASP, 118, 94 Castor, J. I. 1970, MNRAS, 149, 111 Kim, S., Dopita, M. A., Stavely-Smith, L., & Bessell, M. S. 1999, AJ, 118, Dufour, R. J. 1984, in IAU Symp. 108, Structure and Evolution of the Magellanic 2797 Clouds, ed. S. van den Bergh & K. S. de Boer ( Dordrecht: Reidel), 353 Kim, S., Walsh, W., & Xiao, K. 2004, ApJ, 616, 865 Dufour, R. J., Shields, G. A., & Talbot, R. J., Jr. 1982, ApJ, 252, 461 Kohno, K. 2005, in ASP Conf. Ser. 344, The Cool Universe: Observing Cosmic Ezawa, H., Kawabe, R., Kohno, K., & Yamamoto, S. 2004, Proc. SPIE, 5489, Dawn, ed. C. Lidman & D. Alloin (San Francisco: ASP), 242 763 Kutner, M. L., et al. 1997, A&AS, 122, 255 Fukui, Y. 2007, in IAU Symp. 237, Triggered Star Formation in a Turbulent Ladd, N., Purcell, C., Wong, T., & Robertson, S. 2005, Publ. Astron. Soc. ISM, ed. B. G. Elmegreen & J. Palous (Cambridge: Cambridge Univ. Press), Australia, 22, 62 31 MacLaren, I., Richardson, K. M., & Wolfendale, A. W. 1988, ApJ, 333, 821 Fukui, Y., Mizuno, N., Yamaguchi, R., Mizuno, A., & Onishi, T. 2001, PASJ, Mizuno, N., et al. 2001, PASJ, 53, 971 53, L41 Muraoka, K., et al. 2007, PASJ, 59, 43 Fukui, Y., et al. 1999, PASJ, 51, 745 Poglitsch, A., Krabbe, A., Madden, S. C., Nikola, T., Geis, N., Johansson, L. E. B., ———. 2007, ApJ, submitted Stacey, G. J., & Sternberg, A. 1995, ApJ, 454, 293 Garay, G., Johansson, L. E. B., Nyman, L.-8., Booth, R. S., Israel, F. P., Rolleston, W. R. J., Trundle, C., & Dufton, P. L. 2002, A&A, 396, 53 Kutner, M. L., Lequeux, J., & Rubio, M. 2002, A&A, 389, 977 Schöier, F. L., van der Tak, F. F. S., van Dichoeck, E. F., & Black, J. H. 2005, Goldreich, P., & Kwan, J. 1974, ApJ, 189, 441 A&A, 432, 369 Goldsmith, P. F., & Langer, W. D. 1978, ApJ, 222, 881 Skillman, E. D., & Klein, U. 1988, A&A, 199, 61 Heikkilä, A., Johansson, L. E. B., & Olofsson, H. 1999, A&A, 344, 817 Sorai, K., et al. 2000, Proc. SPIE, 4015, 86 Hughes, A., Wong, T., Ekers, R., Staveley-Smith, L., Filipovic, M., Maddison, ———. 2001, ApJ, 551, 794 S., Fukui, Y., & Mizuno, N. 2006, MNRAS, 370, 363 Stacey, G. J., Jaffe, D. T., Geis, N., Genzel, R., Harris, A. I., Poglitsch, A., Ikeda, M., et al. 1999, ApJ, 527, L59 Stutzki, J., & Townes, C. H. 1993, ApJ, 404, 219 Israel, F. P., de Graauw, Th., van de Stadt, H., & de Vries, C. P. 1986, ApJ, 303, Sun, K., Kramer, C., Ossenkopf, V., Bensch, F., Stutzki, J., & Miller, M. 2006, 186 A&A, 451, 539 Israel, F. P., & Koornneef, J. 1979, ApJ, 230, 390 Werner, M. W., Becklin, E. E., Gatley, I., Ellis, M. J., Hyland, A. R., Robinson, G., Israel, F. P., Maloney, P. R., Geis, N., Herrmann, F., Madden, S. C., Poglitsch, & Thomas, J. A. 1978, MNRAS, 184, 365 A., & Stacey, G. J. 1996, ApJ, 465, 738 Williams, J. P., & Blitz, L. 1998, ApJ, 494, 657 Israel, F. P., et al. 2003, A&A, 406, 817 Yamaguchi, R., et al. 2001, PASJ, 53, 985 Johansson, L. E. B., Olofsson, H., Hjalmarson, A., Gredel, R., & Black, J. H. Yonekura, Y., Asayama, S., Kimura, K., Ogawa, H., Kanai, Y., Yamaguchi, N., 1994, A&A, 291, 89 Barnes, P. J., & Fukui, Y. 2005, ApJ, 634, 476 Johansson, L. E. B., et al. 1998, A&A, 331, 857